BACKGROUND: Tidal expiratory flow limitation (EFL) contributes to chronic dyspnea and exercise intolerance in COPD patients. It can be assessed with the negative expiratory pressure (NEP) technique and is expressed as either the percentage of the tidal volume over which EFL occurs (EFL%Vt) or according to more detailed three-point or five-point scoring systems. The aim of this study was to evaluate the reproducibility of the NEP technique in COPD patients. METHODS: Tidal EFL was evaluated with NEP in 18 subjects with stable COPD (FEV(1) range, 18 to 75% predicted) on two occasions (mean retest interval, 8.2 days) by the same rater. Agreement between testing occasions was assessed with the kappa statistic for the 3-point and 5-point EFL scores, and with the coefficient of repeatability for EFL%Vt. RESULTS: On the first testing occasion, nine subjects had no EFL, four subjects had EFL in the supine position, and five subjects had EFL in the sitting and the supine position. Using the 3-point score, agreement was present in 14 of 18 subjects at time 2 (kappa = 0.66), indicating substantial agreement. Using the 5-point score, agreement was seen in 13 of 18 subjects (kappa = 0.61), also indicating substantial agreement. The reproducibility of EFL%Vt measurements was lower than that required to reliably detect clinical change in both the sitting and supine positions (coefficient of repeatability, 37% and 58%, respectively). CONCLUSIONS: The 3-point and 5-point scores provide a reproducible assessment of EFL in COPD patients. The classification of EFL as a percentage of tidal volume is less reproducible, and large changes are required to be confident that real clinical change has occurred.